Show more

社交网络用户生命周期内的语言变化 @mature 

#论文导读 #老奶奶都能懂的论文导读

每个社交网络里面都会演化出特定的语言习惯(linguistic norm),也就是站内的“黑话”。本论文研究了用户在不同生命周期里是如何使用这些黑话的。

论文研究对象是两个啤酒评论社区BeerAdvocate 和 RateBeer 长达十年的数据。十年里,这两个线上社区经历了多轮的用户更迭,论坛的黑话也翻新了好多次。

依靠语言模型,研究人员发现了用户在不同生命周期里面对黑话的反应是不同的:早期加入的时候会积极学习黑话,使自己的语言接近社区的语言。等相似度达到顶峰之后,用户就会变得保守,不再积极接受新的黑话。随着社区黑话不断更新,老用户依然会坚持使用以前的黑话,直到发现自己的黑话已经跟社区新演化出来的黑话对不上了(见左图),这时候就是老用户离开社区的时刻了。令人惊讶的是,这个规律是非常稳定的,虽然用户生命周期有长有短,经过标准化之后,也会呈现出早期对社区黑话开放后期逐渐保守的特征 (见下图Figure 8,右图)。说明使用社区黑话体现了对社区的热爱程度。

根据这些发现,研究人员尝试预测用户的生命周期。只要对比用户的前20条帖子的语言和当前社交网络的总体语言习惯,就能够一定程度上预测该用户是否很快就会离开社交网络。

论文传送门:web.stanford.edu/~jurafsky/pub
DOI:没有,ACM的数据库暂时down掉了

when i heard the phrase "the average iq is below room temperature", i took it as even more hurtful than intended because i assumed room temperature is in celsius instead of fahrenheit

@caasih @ymhuang0808 我覺得我最近像我 NanoVNA 上的史密斯圖一樣,在平面上繞圈圈

gfverif's way of doing quick and dirty verification of existing C code is somewhere between genius and madness - define a new "mockup" integer type and use C++ to overload the operators like plus or minus, instead of doing calculations, it generates SageMath scripts of equivalent operations in algebra. This way, the entire C algorithm can be automatically extracted without writing any parser or compiler... The only problem: conditional branches using data are not supported, but you're not supposed to do that in crypto anyway... 🤣

Show thread

Successfully got a custom SGB border loaded. I will not be accepting artistic criticism at this time due to the fact that it’s already perfect

Show thread

PDP-11's (optional) multiplier can only perform SIGNED multiplication, to do unsigned multiplication in branchless, constant-time cryptography code, I need to do this convoluted sequence of adjustments... Totally a brain teaser. #retrocomputing #PDP11

Show thread

"Icon" even has a demo version that comes with a little demonstration of what this hidden CGA mode is capable of.

Show thread

@pcjustin 哇,我前幾天剛剛補完這部經典作品,結果你就拿到海報了

The entire reason why Nvidia can block crypto mining in the current way is that their drivers are closed source and they can basically add whatever secret sauce they want. Not a good thing.


M 頭 (PL-259) 的中心導體和香蕉插 (Banana Plug) 其實是一樣的,因此香蕉插可以插進去 M 頭母座 (SO-239) 裡面

@miaoski 這不是當然的麼?不過 HackRF 的 ADC 也只有 8 Bit,Dynamic Range 不用抱太大希望

:aru_0560: 以前说xmpp耗电量大、协议冗余大,不适合做移动端IM。但是拿现在的眼光看,它耗电量再大,能比现在内置小程序框架的傻逼IM软件耗电量大吗?至于流量问题,现在的移动端用户都拿流量刷短视频了,会在意文本消息那点流量的冗余?

@miaoski 這種是 Magnetic Loop 沒有極化吧?我講的是指向性,跟極化無關

Show more
Mastodon

The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!